JP2002014700A - Method and device for processing audio signal, and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To speedily turn digital audio signals into high efficient signs. SOLUTION: A mask amount computing section 104 converts frequency analyzed digital audio signals into first and second information. The first information indicates the values located within a prescribed range and the second information indicates the exponent of the power of ten. The section 104 searches a conversion table 109 to obtain the logarithm of the first information. Moreover, the section 104 obtains the sound pressure level of the digital audio signals on the basis of the logarithm of the first information and the second information and computes the mask amount on the basis of the obtained sound pressure level.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for encoding digital audio signals with high efficiency.

[0002]

2. Description of the Related Art As a technique for encoding digital audio signals with high efficiency, there is an MPEG audio encoding method based on ISO / IEC11172-3.

FIG. 5 is a block diagram showing a configuration of a speech encoding device for realizing the MPEG audio encoding method. In the figure, 1 is an input terminal, 2 is a sub-band division unit, 3 is an FFT (fast Fourier transform) operation unit, 4 is a mask amount operation unit, 5 is a quantization step operation unit, 6 is a quantization unit, and 7 is A transmission format forming unit 8 is an output terminal.

[0004] An input terminal 1 inputs a digital audio signal. The sub-band division unit 2 divides the digital audio signal from the input terminal 1 into predetermined time intervals (audio frames), and maps the divided audio frames to a plurality of frequency bands. The FFT operation unit 3 analyzes the frequency of the digital audio signal from the input terminal 1 using the fast Fourier transform (FFT). Mask amount calculation unit 4
Calculates a sound pressure level [dB] from the analysis result of the FFT calculation unit 3, and then calculates a mask amount using the sound pressure level and a predetermined psychoacoustic model. Quantization step calculator 5
Estimates the number of bits to be allocated to each subband based on the mask amount calculated by the mask amount calculation unit 4 (that is, sets a quantization step for each subband). The quantization unit 6
Each sub-band is quantized based on the quantization step obtained by the quantization step calculator 5. The transmission format forming unit 7 converts information such as the output of the quantization unit 6 (a highly efficient encoded digital audio signal) and the quantization step obtained by the quantization step calculation unit 5 into a predetermined transmission format. The output terminal 8 outputs a highly efficient encoded digital audio signal.

[0005]

In the above-mentioned MPEG audio coding system, when a sound pressure level is obtained from the analysis result of the FFT operation unit 3, a logarithmic operation represented by the equation (1) must be executed. X (k) = 10Log (FFT (k)) (1)

Here, FFT (k) is the analysis result of the FFT operation unit 103 at the frequency position k, and X (k) is the sound pressure level (decibel amount) at the frequency position k. When the above-mentioned MPEG audio encoding system is realized by software, such a logarithmic operation must be processed by software for each frequency position k. However,
When all such logarithmic operations are processed by software, the amount of calculation and the processing time increase, and there is a problem that real-time processing becomes difficult.

The present invention has been made in view of the above problems, and has as its object to provide a technique for encoding digital audio signals at high speed with high efficiency.

[0008]

In order to achieve the object of the present invention, an audio signal processing method according to the present invention comprises a converter for converting a frequency-analyzed digital audio signal into first information and second information. A step of searching a table that holds a logarithm of a plurality of values within a predetermined range to obtain a logarithm of the first information; and searching for a logarithm of the first information and a logarithm of the first information and the second information. Calculating a sound pressure level of the digital audio signal based on the calculated sound pressure level, and calculating a mask amount based on the calculated sound pressure level.

Further, the audio signal processing apparatus of the present invention comprises a conversion means for converting a frequency-analyzed digital audio signal into first information and second information, and a logarithm of a plurality of values within a predetermined range. Search for a table that holds
Searching means for calculating the logarithm of the information, calculating the sound pressure level of the digital audio signal based on the logarithm of the first information and the second information, and calculating the mask amount based on the calculated sound pressure level And a mask amount calculating means.

Further, the storage medium of the present invention is characterized by holding a program code for realizing the above-mentioned audio signal processing method in a computer.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below in detail with reference to the accompanying drawings.

FIG. 1 is a diagram showing a configuration of a speech coding apparatus according to an embodiment of the present invention. In the present embodiment,
The MPEG audio encoding method (ISO / IEC), which is one of the techniques for encoding digital audio signals with high efficiency
11172-3) will be described. The speech encoding device according to the present embodiment can be partially or entirely realized by software. In this case, the control unit (CP
U and MPU) read a program for realizing this software from the storage medium, and execute various processes according to the read program.

In FIG. 1, reference numeral 101 denotes an input terminal;
Is a subband division unit, 103 is an FFT (fast Fourier transform) operation unit, 104 is a mask amount operation unit, 105 is a quantization step operation unit, 106 is a quantization unit, 107 is a transmission format formation unit, 108 is an output terminal, 109 is a conversion table.

An input terminal 101 inputs a digital audio signal, and supplies the input digital audio signal to a subband division unit 102 and an FFT operation unit 103.
The sub-band dividing unit 102 divides the digital audio signal into predetermined time intervals (audio frames), and divides the divided audio frames into frequency bands (sub-bands) in units of 750 Hz using an overlapping polyphase filter bank (OPFB). Band). The data of each subband is quantized by the quantization unit 106 later.

On the other hand, the FFT operation unit 103 analyzes the frequency of the digital audio signal from the input terminal 101 using the fast Fourier transform. This analysis result is supplied to the mask amount calculation unit 104. The mask amount calculation unit 104
After obtaining the sound pressure level [dB] from the analysis result of the FFT calculation unit 103 using the conversion table 109, the mask amount is calculated using the sound pressure level and a predetermined psychoacoustic model. This mask amount is supplied to the quantization step calculation unit 105.

The quantization step calculator 105 adaptively estimates the number of bits to be allocated to each subband based on the mask amount obtained by the mask amount calculator 104 (ie, adaptively determines the quantization step of each subband). Set).

The quantization unit 106 is a quantization unit that quantizes each subband based on the quantization step obtained by the quantization step calculation unit 105. The transmission format forming unit 107 outputs the output of the quantizing unit 106 (a highly efficient encoded digital audio signal) and the quantization step calculating unit 10
The transmission format forming unit converts the information such as the quantization step obtained in step 5 into a predetermined transmission format. The output terminal 108 outputs a highly efficient encoded digital audio signal.

FIG. 2 is a flowchart illustrating an algorithm for creating the conversion table 109. The conversion table 109 holds logarithms of a plurality of values existing within a predetermined range. In the present embodiment, an example in which the predetermined range is set to 1 or more and less than 10 in consideration of log periodicity will be described. Also, in the present embodiment, considering that the significant digits of the minimum audible limit value are two decimal places in decibels, the logarithm of the value increasing by 0.01 from 1.00 is calculated and converted. An example of storing in the table 109 will be described.

In step S201, the variable seed is set to 1
Set 00. In step S202, the variable seed
Multiply d by 0.01 and calculate the logarithm of the result of this multiplication.

In step S203, step S202
The logarithm TBLexp [seed] obtained in step (1) is stored in the conversion table 109 at an address corresponding to the variable seed. Note that the address of the conversion table 109 is the variable se
Like ed, it changes from 100 to 999.

In step S204, the variable seed is set to 1
000 is determined. If not equal, the process proceeds to step S205, 1 is added to the variable seed, and the process proceeds to step S201. On the other hand, if they are equal,
The process ends.

With the above configuration, the conversion table 109 can store logarithms from 1.00 to 9.99.

FIG. 3 is a diagram showing a configuration of the mask amount calculation unit 104 according to the present embodiment. In the present embodiment,
A case where the mask amount calculation unit 104 is realized by software will be described.

When calculating the sound pressure level from the analysis result of the FFT operation unit 103, the mask amount operation unit 104 calculates the expression (1)
The operation shown in must be performed. X (k) = 10Log (FFT (k)) (1)

Here, FFT (k) is the analysis result of the FFT operation unit 103 at the frequency position k, and X (k) is the sound pressure level (decibel amount) at the frequency position k. In the mask amount calculation unit 104 of the present embodiment, the sound pressure level X (k) is obtained as follows in order to perform the calculation represented by Expression (1) at high speed. That is, instead of directly calculating the logarithm of FFT (k), FFT (k) is once converted to “A × 10 ^B ”, and the FFT (k) is calculated based on the sum of the logarithm of “A” and “B”.
The logarithm of (k) is found, and the sound pressure level X (k) is found. Here, “A” indicates a value that exists within a predetermined range (ie, 1 or more and less than 10) set when the conversion table 109 is created, and “B” indicates a power of 10.

In FIG. 3, reference numeral 301 denotes a conversion module;
302 is a table search module, 303 is a decibel value calculation module, and 304 is a mask amount calculation module.

The conversion module 301 includes an FFT operation unit 1
FFT (k), which is the analysis result of 03, is converted into “A × 10 ^B ”. “A” obtained here is supplied to the table search module 302 as first information, and “B” is supplied to the decibel value calculation module 303 as second information.

The table search module 302 searches the conversion table 109 and obtains the logarithm of the first information (that is, "A") obtained by the conversion module 301.

The decibel value calculation module 303 has a first
Of the information (ie, “A”) and the second information (ie,
(B)) to calculate the sound pressure level X (k).

The mask amount calculation module 304 calculates the sound pressure level X (k) obtained by the decibel value calculation module 303.
And a predetermined psychoacoustic model to calculate the mask amount. This mask amount is supplied to the quantization step calculation unit 105.

FIG. 4 shows an FF using the conversion table 109.
5 is a flowchart illustrating an algorithm for obtaining a mask amount from an analysis result of a T operation unit 103.

In step S 401, the transform module 301 executes the FFT
Enter (k). Here, when the conversion table 109 is created from values to the second decimal place, the conversion module 301 rounds the FFT (k) to the second decimal place by either rounding down, rounding up, or rounding.
In step S402, the conversion module 301
Reset the variable const to 0. Note that the variable con
st corresponds to “B” described above.

In step S403, the conversion module 301 determines whether FFT (k) is less than 10.0. If FFT (k) is 10.0 or more, the process proceeds to step S404. If FFT (k) is less than 10.0, the process proceeds to step S407.

In step S404, 1 is added to the variable const. In step S405, FFT (k) is multiplied by 0.1. In step S406, step S4
It is determined whether the multiplication result obtained in step 05 is less than 10.0. If FFT (k) is 10.0 or more, the process returns to step S404. If FFT (k) is less than 10.0, the process proceeds to step S411.

As described above, when the FFT (k) is 10.0 or more, the conversion module 301 determines that the FFT (k) is 1
Multiply FFT (k) by 0.1 until it is less than 0.0,
Add 1 to the variable const.

In step S407, the conversion module 301 determines whether FFT (k) is less than 1.00. If FFT (k) is less than 1.00, the process proceeds to step S408. On the other hand, if FFT (k) is equal to or greater than 1.00, FFT (k) is an existing value within the range of 1 to less than 10, and the process proceeds to step S411.

In step S408, 1 is subtracted from the variable const. In step S409, FFT (k)
Multiply by 10. In step S410, step S4
It is determined whether the multiplication result obtained in step 09 is less than 1.00. If FFT (k) is less than 1.00, the process returns to step S408; otherwise, the process proceeds to step S411.

Thus, if FFT (k) is less than 1.00, transform module 301 multiplies FFT (k) by 10 until FFT (k) is greater than or equal to 1.00,
Subtract 1 from the variable const.

In step S411, the table search module 302 obtains an address adr corresponding to FFT (k) (corresponding to "A" described above) converted to a value of 1 or more and less than 10. Specifically, the address adr is obtained from the integer part of the result of multiplying FFT (k) by 100. Then, the table search module 302 searches the conversion table 109 based on the obtained address adr, and finds the logarithmic TBLexp [ad held by the address adr.
r] (corresponding to the logarithm of “A”).

In step S412, the decibel value calculation module 303 adds the logarithmic TBLexp [adr] obtained in step S411 and the variable const, and
The sound pressure level X (k) is obtained by multiplying the addition result by 10. That is, the logarithmic TBLexp obtained in step S411
The logarithmic operation shown in Expression (1) is realized based on [adr] and the variable const.

When calculating the sound pressure level X (k) from the FFT (k) at the next frequency position k, the mask value calculating module 304 proceeds to step S401.
The mask value calculation module 304 proceeds to step S414.

In step S414, the mask value calculation module 304 calculates a mask amount based on the sound pressure level X (k) obtained in step S412. The mask amount thus obtained is supplied to the quantization step calculation unit 105. The quantization step calculation unit 105 adaptively sets the quantization step of each subband based on the mask amount.

As described above, according to the present embodiment, FFT (k) is not directly logarithmically calculated,
Since T (k) is once converted to “A × 10 ^B ”, and the logarithm of FFT (k) is obtained based on the sum of the logarithm of “A” and “B”, the logarithmic calculation is more efficient than the usual logarithmic calculation.

The sound pressure level X (k) can be calculated at high speed, and the digital audio signal can be encoded at high speed with high efficiency. Further, according to the present embodiment, since conversion table 109 is created in consideration of the significant figures of the minimum audible limit value, it is possible to reduce the deterioration of the calculation accuracy.

In the present embodiment, a case has been described where the logarithms of the values up to the second decimal place are calculated and stored in the conversion table 109, but the present invention is not limited to this. The logarithms of the values up to the third decimal place may be obtained and stored in the conversion table 109.

In this embodiment, the case where the above-described exponent analysis algorithm is applied to the MPEG audio coding system conforming to ISO / IEC11172-3 has been described, but the present invention is not limited to this. ATRAC
(Adaptive Transform Acoustic Coding) method, PASC (Precision Adaptive Subband Coding) method and the like can be applied to a speech coding method using psychoacoustic characteristics.

[0047]

[Other Embodiments] The present embodiment can be applied to a system including a plurality of devices.

In the present embodiment, as described above, a storage medium storing software program codes for realizing the functions of the present embodiment is supplied to a system or an apparatus, and a control unit (control unit) of the system or the apparatus is provided. Needless to say, the present invention is also achieved when a CPU (including a CPU and an MPU) reads out and executes the program code stored in the storage medium.

In this case, the program code itself read from the storage medium implements the functions of the above-described embodiment, and the storage medium storing the program code constitutes the present invention. When the computer executes the readout program codes, not only the functions of the above-described embodiments are realized, but also an operating system (OS) running on the computer based on the instructions of the program codes. It goes without saying that a case where some or all of the actual processing is performed and the functions of the above-described embodiments are realized by the processing is also included.

Further, after the program code read from the storage medium is written into a memory provided in a function expansion card inserted into the computer or a function expansion unit connected to the computer, the program code is read based on the instruction of the program code. , The CPU provided in the function expansion card or the function expansion unit performs part or all of the actual processing,
It goes without saying that a case where the function of the above-described embodiment is realized by the processing is also included.

[0051]

As described above, according to the present invention, digital audio signals can be encoded at high speed with high efficiency.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a speech encoding device according to the present embodiment.

FIG. 2 is a flowchart illustrating a procedure for creating a conversion table in the present embodiment.

FIG. 3 is a block diagram illustrating a configuration of a mask amount calculation unit according to the present embodiment.

FIG. 4 is a flowchart illustrating a processing procedure of a mask amount calculation unit according to the present embodiment.

FIG. 5 is a block diagram illustrating a configuration of a conventional speech encoding device.

Claims (13)

[Claims] A conversion step of converting a frequency-analyzed digital audio signal into first information and second information; and searching a table holding a logarithm of a plurality of values within a predetermined range, A search step of obtaining a logarithm of the first information; obtaining a sound pressure level of the digital audio signal based on the logarithm of the first information and the second information; and a mask amount based on the obtained sound pressure level. And calculating a mask amount. 2. The audio signal processing method according to claim 1, wherein the predetermined range is 1 or more and less than 10. 3. The converting step includes: converting a frequency-analyzed digital audio signal to a value within the predetermined range;
2. The method according to claim 1, wherein the data is converted into a product of a power of zero.
Or the audio signal processing method according to 2. 4. The audio signal processing method according to claim 1, wherein the first information indicates a value within the predetermined range. 5. The audio signal processing method according to claim 1, wherein the second information indicates an exponent of a power of 10. 6. The mask amount calculating step, wherein a sound pressure level of the digital audio signal is obtained based on a result obtained by adding a logarithm of the first information and the second information. The audio signal processing method according to any one of claims 1 to 5. 7. A conversion means for converting the frequency-analyzed digital audio signal into first information and second information, and searching a table holding a logarithm of a plurality of values within a predetermined range, Searching means for obtaining the logarithm of the first information; obtaining the sound pressure level of the digital audio signal based on the logarithm of the first information and the second information; An audio signal processing device comprising: 8. The audio signal processing device according to claim 7, wherein the predetermined range is 1 or more and less than 10. 9. The converting means converts the frequency-analyzed digital audio signal to a value within the predetermined range and 1
8. The method according to claim 7, wherein the data is converted to a product of a power of zero.
Or the audio signal processing device according to 8. 10. The audio signal processing apparatus according to claim 7, wherein the first information indicates a value within the predetermined range. 11. The audio signal processing apparatus according to claim 7, wherein the second information indicates an exponent of a power of 10. 12. The apparatus according to claim 1, wherein the mask amount calculating means obtains a sound pressure level of the digital audio signal based on a result obtained by adding a logarithm of the first information and the second information. The audio signal processing device according to any one of 7 to 11. 13. A storage medium storing program codes for realizing the audio signal processing method according to claim 1. Description: